UAV obstacle avoidance using image processing techniques

Novel algorithms were developed for use on a low-price, consumer four-rotor helicopter with limited on-board computational capacity. Simplified occupancy grid maps, basic wall-following and obstacle avoidance were constructed from laser rangefinder data using image processing primitives in OpenCV. These algorithms generated obstacle free paths along walls in preliminary testing. The resulting system will allow untrained and unskilled users to teleoperate a low-cost flight platform, assembled from commercially available parts for remote exploration in quasi-static indoor environments.